See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12814766

Microanalysis of cardiolipin in small biopsies including skeletal muscle from

patients with mitochondrial disease

Article  in  Journal of Lipid Research · October 1999

DOI: 10.1016/S0022-2275(20)33404-0 · Source: PubMed

CITATIONS READS

79 166

7 authors, including:

Stephen Doty Thomas J J Blanck

Hospital for Special Surgery New York University
244 PUBLICATIONS   12,303 CITATIONS    177 PUBLICATIONS   3,363 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

FT-IRIS View project

Iron therapy and bone health in Chronic Kidney Disease View project

All content following this page was uploaded by Stephen Doty on 01 June 2014.

The user has requested enhancement of the downloaded file.

 Microanalysis of cardiolipin in small biopsies including
skeletal muscle from patients with mitochondrial disease
Michael Schlame,* Sara Shanske,** Steven Doty,† Thomas König,* Thomas Sculco,§
Salvatore DiMauro,** and Thomas J. J. Blanck1,*
Departments of Anesthesiology,* Pathology,† and Orthopedic Surgery,§ Hospital for Special Surgery,
Cornell University Medical College, 535 E. 70th Street, New York, NY, and Department of Neurology,**
College of Physicians and Surgeons, Columbia University, New York, NY
Abstract Cardiolipin is a specific mitochondrial phospho-
lipid that is present in mammalian tissues in low concentra-
tion. To measure cardiolipin in small biopsies from patients
with mitochondrial disease, we developed a new technique
that can detect subnanomolar levels of well-resolved molec-
ular species, the most abundant of which are tetralinoleoyl-
cardiolipin (L4) and trilinoleoyl-oleoyl-cardiolipin (L3O). To
this end, a fluorescence-labeled derivative of cardiolipin (2-
[naphthyl-19-acetyl]-cardiolipin dimethyl ester) was formed
and analyzed by high performance liquid chromatography.
Cardiolipin was measured in skeletal muscle biopsies from
8 patients with mitochondrial disease and in 17 control sub-
jects. In 5 patients with mitochondrial disease, cardiolipin
content was higher than normal (2.4–7.0 vs. 0.4–2.2 nmol/
mg protein). In 3 patients with mitochondrial disease, the
L4/L3O ratio was lower than normal (2– 4 vs. 4– 6). Cardio-
lipin was also measured in various rat and dog muscle tis-
sues. The L4/L3O ratio was higher in condensed “muscle”
type mitochondria (heart ventricle, skeletal muscle, ratios
4–7) than in orthodox “liver” type mitochondria (liver,
smooth muscle, heart auricular appendage, H9c2 myo-
blasts, ratios 0.4–3), suggesting that the L4/L3O proportion
is important for cristae membrane structure. We con-
cluded that the L4/L3O ratio is a tissue-specific variable that
may change in the presence of mitochondrial disease. The
new method is suitable to measure cardiolipin in muscle
biopsies in order to estimate concentration of mitochon-
dria.—Schlame, M., S. Shanske, S. Doty, T. König, T. Sculco,
S. DiMauro, and T. J. J. Blanck. Microanalysis of cardiolipin
is small biopsies including skeletal muscle from patients
with mitochondrial disease. J. Lipid Res. 1999. 40: 1585–
1592.
Supplementary key words fluorescence • lipid analysis • high perfor-
mance liquid chromatography • mitochondrial DNA • neuromuscular
disease
Cardiolipin is an acidic phospholipid with a unique
dimeric structure including four fatty acids (1). It is found
in both bacteria and mitochondria. While bacteria express
variable amounts of cardiolipin, depending on the stage
of growth, mitochondria contain cardiolipin as an essen-
tial component of their cristae membrane (2 –5). Until re-
cently it was believed that mitochondria are absolutely de-
pendent on cardiolipin. However, a cardiolipin synthase
null mutant in Saccharomyces cerevisiae was viable (6), prob-
Downloaded from www.jlr.org by guest, on June 7, 2013
ably because phosphatidylglycerol can replace cardiolipin
(7). Although the exact biological function of cardiolipin
still remains to be defined, its high affinity to various
membrane proteins suggests involvement in the intricate
structure of the mitochondrial cristae membrane (3 –5).
Indeed, a mammalian cell line made deficient in cardio-
lipin showed markedly altered mitochondrial ultrastruc-
ture (8).
Cardiolipin was first discovered in beef heart tissue (9).
Soon it was recognized that it was not specific to the heart
but it was merely more abundant there because of the
high concentration of mitochondria in cardiac myocytes.
In tissues that contain less mitochondria, it has been noto-
riously difficult to measure cardiolipin by conventional
lipid analytical techniques. This has become a problem
because there has been renewed interest in cardiolipin as
a potential factor in several pathologies, such as thyroid
disease (10–12), adriamycin toxicity (13), free radical-
mediated disorders (14), complement activation (15), as
well as in aging (12, 16). Some progress in the microdetec-
tion of cardiolipin has been made by using the fluorescent
dye 10-N-nonyl acridine orange, which displays specificity
for cardiolipin upon binding to intact membranes (17).
To date, thyroid dysfunction is the only disease that has
been unequivocally linked to changes in cardiolipin con-
tent. While the hypothyroid state leads to a decrease of
cardiolipin (18), the opposite is true for hyperthyroidism
(19), and a direct effect of thyroxin on the biosynthetic
enzymes of cardiolipin was reported (10, 20). As thyroxin
Abbreviations: HPLC, high performance liquid chromatography;
L4, tetralinoleoyl-cardiolipin; L3O, trilinoleoyl-oleoyl-cardiolipin; L2O2,
dilinoleoyl-dioleoyl-cardiolipin; MELAS, multifocal encephalomalacia,
lactic acidosis, and stroke-like episodes.
1 To whom correspondence should be addressed.
Journal of Lipid Research Volume 40, 1999 1585
is a well-known stimulator of mitochondrial biogenesis, it
may be hypothesized that cardiolipin has some role in the
formation of mitochondria. It should be noted that mam-
malian cardiolipin displays an unusual preference for li-
noleic acid (2–4, 21). As this preference is not shared by
other mammalian phospholipids, it can be inferred that
the specific acyl pattern of cardiolipin is important for its
function in the mitochondria. Therefore, analysis of car-
diolipin in pathological tissues ought to measure both
cardiolipin content and its pattern of molecular species.
To study cardiolipin in muscle biopsies from children with
mitochondrial disease, we developed a novel technique to
quantitate cardiolipin in small tissue samples. The method
exploits the fact that the acid form of cardiolipin can be
methylated yielding an unusually apolar phospholipid,
which can be easily separated from normal phospholipids
and is very suitable for reversed-phase HPLC. The methy-
lated cardiolipin was labeled with the naphthyl moiety to
allow HPLC with fluorescence detection. The chromatog-
raphy resolved molecular species of cardiolipin.
MATERIALS AND METHODS
Patients
Skeletal muscle biopsies were obtained for diagnostic pur-
poses from adults (age 18–74 years) and children (age 1–12
years). Additional muscle samples were obtained from patients
who underwent knee surgery using a protocol approved by the
institutional review board of the Hospital for Special Surgery. All
patients gave written informed consent. Patients included eight
subjects with well-documented mitochondrial disorders: four
children (three males, one female, ages 1 to 9 years) with gener-
alized cytochrome oxidase deficiency, two women (ages 18 and
35 years) with MELAS 3243 (A3243G point mutation in the
tRNALeu(UUR) gene of mitochondrial DNA, producing a syndrome
characterized by multifocal encephalomalacia, lactic acidosis,
and strokelike episodes), one child (female, 1 year) with Leigh
syndrome (T8993G mutation of mitochondrial DNA), and one
man (age 34 years) with multiple deletions of mitochondrial
DNA. The latter case has been reported before (22).
Materials
Muscle tissues were obtained from male Sprague-Dawley rats
(skeletal, smooth, and heart muscles) and female beagles (heart
muscles). Rat skeletal muscle tissue was obtained from the dia-
phragm. Rat smooth muscle was taken from the esophagus. The
animals were anesthetized with isoflurane in a procedure ap-
proved by the Institutional Animal Care and Use Committee of
the Hospital for Special Surgery. Tissues were homogenized in 50
mm Tris-buffered 0.15 m KCl (pH 7.5) using first a Polytron
PT300 tissue grinder and then a Potter-Elvehjem homogenizer
(100 mg tissue wet weight per ml buffer). The H9c2 myoblast cell
line was obtained from the American Type Culture Collection
(Rockville, MD). Cells were grown in Dulbecco’s modified Eagle’s
medium, containing 10% fetal calf serum, harvested by trypsin di-
gestion, collected by centrifugation, and resuspended in buffered
saline prior to lipid extraction. Tetrastearoyl-cardiolipin was pro-
duced by hydrogenation of commercial bovine heart cardiolipin
(Sigma, St. Louis, MO). The commercial cardiolipin (2.5 mg)
was dried under a stream of nitrogen and redissolved in 1.0 ml
acetonitrile–ethanol 1:1 (by volume). Platinium(IV) oxide (25
mg) was added and the suspension was bubbled with hydrogen
1586 Journal of Lipid Research Volume 40, 1999
gas for 30 min. The catalyst was removed by centrifugation and
filtration, the clear solution was dried under nitrogen and redis-
solved in ethanol–chloroform 2:1 (by volume). Supelclean solid-
phase extraction tubes were obtained from Supelco (Bellefonte,
PA) and the C18-Hypersil HPLC column was from Sigma (St.
Louis, MO). Naphthylacetic anhydride was obtained from Aldrich
(Milwaukee, WI). All other chemicals were of analytical grade.
Quantitative analysis of cardiolipin
Lipids were extracted from muscle tissue (50–100 mg wet
weight per sample) or cell homogenates (2–4 3 106 cells per
sample) according to Bligh and Dyer (23). Tetrastearoyl-cardio-
lipin (3.6 nmol) was added as an internal standard at the initial
step of lipid extraction. After extraction, lipids were dried under
nitrogen and acidified by the following method. Ice-cold metha-
nol (2 ml), chloroform (1 ml), and 0.1 m HCl (1 ml) were
added, the sample was vortexed and incubated on ice for 5 min.
Phase separation was achieved by addition of 1 ml of chloroform
and 1 ml of 0.1 m HCl and the chloroform-phase was recovered,
followed by re-extraction with 2 ml of chloroform. The chloro-
form extract was dried under nitrogen and the lipids were
treated with diazomethane. To this end, diazomethane was re-
leased from 1 g of N-methyl-N-nitroso-p-toluenesulfonamide by
addition of 2 ml of ethanol and 0.3 ml of 16 m KOH, and the gas
was trapped in 16 ml of ice-cold chloroform. One milliliter of
Downloaded from www.jlr.org by guest, on June 7, 2013
the chloroform solution, containing diazomethane, was added
to the dry lipid residue and incubated on ice for 15 min. After
that, the solvent was evaporated and the methylated lipids were
redissolved in 0.2 ml of chloroform. This solution was purified
by solid-phase extraction on Supelclean LC-Si tubes (100 mg sil-
ica per tube) equilibrated with diethylether–ethanol 9:1 (by vol-
ume). After loading the sample, the column was eluted with
2 ml diethylether–ethanol 9:1. The eluate was dried under ni-
trogen and then redissolved in a reagent consisting of 25 mmol
1-naphthylacetic anhydride and 25 mmol N,N-dimethyl-4-
aminopyridine dissolved in 250 ml anhydrous pyridine. This mixture
was incubated for 2 h at 40 8C. Then, 6 ml of n-hexane was
added, the sample was spun to sediment non-soluble material,
and the supernatant was further purified by extraction with
water. The hexane-phase was treated with anhydrous sodium sul-
fate to remove traces of water and then evaporated to dryness.
The dry residue was dissolved in 1 ml of n-hexane–diethylether
1:1 (by volume) and subjected to another solid-phase extraction
step. Supelclean LC-Si tubes (100 mg silica) were equilibrated in
n-hexane–diethylether 1:1, the sample was loaded onto the col-
umn and eluted with another 2 ml of n-hexane–diethylether
1:1. Next, the column was eluted with 2 ml of diethylether–
ethanol 95:5 (by volume) and this eluate was collected, dried,
and redissolved in 0.1 ml of n-hexane–ethanol 1:1 (by volume).
Twenty microliters of this solution was separated by HPLC using
a C18-Hypersil (5 mm) column (150 3 3.2 mm). A solvent gradi-
ent was run from acetonitrile–2-propanol 8:2 (by volume) to
acetonitrile–2-propanol 5:5 in 30 min. The HPLC was a Shi-
madzu system consisting of two LC-10AT solvent delivery systems
(total flow rate 1.0 ml/min), a Rheodyne manual 20-ml loop in-
jector, and RF-10AXL fluorescence detector (excitation wave-
length 280 nm, emission wavelength 360 nm, gain 5 2, sensitiv-
ity 5 2). Data were collected and processed by the Shimadzu
CLASS VP software running on a desktop computer. For peak
identification, fractions were collected and processed for mea-
surement of the fatty acid profile.
Other methods
For electron microscopy, muscle samples were dissected into
2–4 mm cubes and placed into cold fixative for 12–18 h. The fix-
ative consisted of 2% paraformaldehyde plus 0.5% glutaral-
dehyde in 0.5 m cacolydate buffer (pH 7.2). Tissue samples were
dehydrated through a graded alcohol series and embedded in
Spurr’s resin. Thin sections were collected and stained with Rey-
nolds lead stain and alcoholic uranyl acetate. Micrographs were
taken at 13,000 original amplification for all samples, using a
Philips CM-12 transmission electron microscope.
Protein concentration was determined according to Lowry et
al. (24). Ultraviolet absorbance spectra were recorded with a
Beckman DU-7400 spectrophotometer. Fluorescence spectra
were recorded with a SLM-Aminco fluorescence photometer.
Fatty acids were measured by gas chromatographic analysis of
their methyl esters (25), using a Supelcowax 10 capillary column
(Supelco, Bellefonte, PA) installed in a GC-17A Shimadzu gas
chromatograph. A temperature gradient was run from 180 to
240 8C at 5 8C/min. Citrate synthase activity was determined spec-
trophotometrically as described by Srere (26). One enzyme unit
was defined as 1 mmol citrate formed per minute. Phospholipid
concentration was determined by colorometric measurement of
phosphate liberated by ashing (27). For statistical presentation,
means are given with standard deviation. Patient groups were
compared by the Mann-Whitney rank sum test.
RESULTS
Microanalysis of cardiolipin
We syntheszied a new chemical derivative of cardiolipin
in order to develop a sensitive and quantitative assay
based on HPLC with fluorescence detection. To this end,
a naphthyl-19-acetyl moiety was linked to the only free hy-
droxyl group of dimethylcardiolipin. The product showed
the characteristic absorbance pattern of the naphthyl
group as well as a fluorescence signal centered around
367 nm when excited at 280 nm (Fig. 1).
For the purpose of quantitative determination of cardi-
olipin in biological extracts, the chemical derivatization
was combined with a step-by-step purification designed to
eliminate other lipids and to isolate the cardiolipin deriva-
tive. First, derivatization with diazomethane was performed
on the whole lipid mixture. This reaction converted only
acidic phospholipids to their methyl esters. In the subse-
quent solid phase extraction, phospholipid methyl esters

Downloaded from www.jlr.org by guest, on June 7, 2013

Fig. 1. Spectral properties of the naphthyl derivative of cardio-
lipin. Spectra of 2-(naphthyl-19-acetyl)-cardiolipin dimethyl ester
were recorded in n-hexane. A: Characteristic segment of the absor-
bance spectrum showing peaks at 272, 282, and 292 nm. B: The fluo-
rescence spectrum, recorded at excitation wavelength of 280 nm,
shows an emission maximum at 367 nm. The right panel shows a
molecular model of cardiolipin (21) with the chemical modifi-
cations made by the derivatization procedure (black, carbon; blue,
hydrogen; red, oxygen; orange, phosphorus).

Schlame et al. Muscle cardiolipin 1587

were sufficiently non-polar to be eluted from silica by
diethylether–ethanol 9:1. Thus, lipid separation after
methylation was associated with considerable purification.
Then, the naphthyl-19-acetyl group was attached and the
sample was further purified to make it suitable for analyti-
cal HPLC. The gradual purification of derivatized cardio-
lipin was documented by thin-layer chromatography (data
not shown).
Analytical HPLC of 2-(naphthyl-19-acetyl-)cardiolipin
dimethyl ester, derived from human skeletal muscle, is
shown in Fig. 2. Four cardiolipin peaks were obtained,
representing four different molecular species. The fifth
peak was the internal standard, a synthetic cardiolipin spe-
cies that does not occur in nature. The time-integral of Fig. 3. Quantitative and qualitative analysis of cardiolipin. A: Com-
the fluorescence peak was proportional to the amount mercial cardiolipin (bovine heart) was subjected to the analytical
of cardiolipin (Fig. 3A), so cardiolipin concentrations could procedure. The committed amount of cardiolipin was plotted
be calculated in reference to the internal standard. The re- against the total peak area in the fluorescence chromatogram. B: Re-
tention time of molecular species is expected to increase tention time of molecular species of partially hydrogenated cardio-
lipin from bovine heart. All species contain four C 18-chains and a
with increasing length of the acyl chains and decrease with
variable number of double bonds.
increasing degree of unsaturation. Accordingly, we found
the logarithmic retention time to be inversely proportional
to the total number of double bonds in all-C18 species of
cardiolipin (Fig. 3B). method described above (Fig. 4). The content was highest

Downloaded from www.jlr.org by guest, on June 7, 2013

Cardiolipin from animal muscle tissues
Cardiolipin content was measured in various muscle tis-
sues from rat and dog, using the fluorescence–HPLC
in heart muscle, much lower in skeletal muscle, and low-
est in smooth muscle. In heart, the ventricle contained
significantly more cardiolipin than the auricular append-
age. The cardiolipin content of cultured H9c2 myoblasts
was also measured. The cell culture contained 2.9 6 0.8
nmol cardiolipin per milligram protein (n 5 3), about as
much as skeletal muscle tissue. This corresponded to 2.2 6
0.6 fmol of cardiolipin per single cell.

Fig. 2. HPLC of cardiolipin from human skeletal muscle. Cardio-

lipin was extracted from a skeletal muscle biopsy of a child with cyto-
chrome oxidase deficiency. The HPLC chromatogram shows the
fluorescence signal, recorded on a voltage scale, versus the reten-
tion time. Peaks represent molecular species of muscle cardiolipin:
trilinoleoyl-linolenoyl-cardiolipin (peak 1), tetralinoleoyl-cardio- Fig. 4. Cardiolipin content of various muscle tissues. Lipid ex-
lipin (L 4, peak 2), trilinoleoyl-oleoyl-cardiolipin (L 3O, peak 3), and tracts from muscle tissues were derivatized, purified, and analyzed
dilinoleoyl-dioleoyl-cardiolipin (L 2O2, peak 4). The internal stan- by HPLC with fluorescence detection. For each tissue, data were
dard (tetrastearoyl-cardiolipin) is in peak 5. obtained from three independent samples.

1588 Journal of Lipid Research Volume 40, 1999


Fig, 5. HPLC of cardiolipin from rat heart. The figure shows the
chromatogram of cardiolipin extracted from the left ventricle and
from the auricular appendages. L 4: tetralinoleoyl-cardiolipin, L 3O:
trilinoleoyl-oleoyl-cardiolipin, L 2O2: dilinoleoyl-dioleoyl-cardiolipin.
The two most abundant cardiolipin species in muscle
tissues were L 4 and L 3O, together accounting for more
than 80% of total cardiolipin. In rat heart, there was a re-
markable difference between cardiolipins extracted from
ventricle and auricular appendage (Fig. 5). Whereas ven-
tricle cardiolipin was predominantly L4, the appendage
contained L 4 and L 3O in about equal amounts. We found
Fig. 6. L 4/L 3O ratio in various muscle tissues. Cardiolipins from
various muscle tissues and from H9c2 myoblast cell culture were an-
alyzed and the ratios of tetralinoleoyl-cardiolipin (L 4 ) to trilino-
leoyl-oleoyl-cardiolipin (L 3O) were determined.
the L 4/L 3O ratio to be a characteristic feature of each
type of muscle (Fig. 6). In rat, the ratio varied from values
higher than 5 to values lower than 0.5. Muscle tissues with
high L 4/L 3O ratio, such as heart ventricle and skeletal
muscle, contained mitochondria with a very regular ar-
rangement of stacked cristae membranes (Fig. 7, A and
C). This is a well-known characteristic of heart and skele-
tal muscle mitochondria. However, in tissues with low
L 4/L 3 O ratio, such as smooth muscle and myoblasts, there
was less cristae alignment (Fig. 7, D and E). Even auricular
mitochondria (Fig. 7 B), although being derived from
heart, had wider cristae sacks with some irregularities,
such that they did not achieve the same kind of stacking
seen in ventricle and skeletal muscle mitochondria.
Skeletal muscle cardiolipin from patients
with mitochondrial disease
Cardiolipin was measured in skeletal muscle biopsies
from patients with mitochondrial disease and in control
subjects (Fig. 8). There was no significant difference be-
tween control biopsies from five children and twelve
adults. In the disease group, five out of eight patients (two
Downloaded from www.jlr.org by guest, on June 7, 2013
patients with cytochrome oxidase deficiency, two patients
with MELAS, and one patient with Leigh syndrome) had
elevated cardiolipin content in skeletal muscle. The differ-
ence in cardiolipin content between seventeen control
subjects and eight patients with mitochondrial disease was
statistically significant (P 5 0.00326).
The pattern of cardiolipin species in skeletal muscles
from patients with mitochondrial disease is shown in
Table 1. L 4 was the predominant species in all biopsies.
The L 4/L3 O ratios were 4 to 6 in control biopsies and in
biopsies from patients with cytochrome oxidase defi-
ciency. However, in two patients with MELAS and in one
patient with multiple deletions of the mitochondrial DNA,
the proportion of L 4-cardiolipin was significantly lower in
favor of more saturated species (L 4/L 3O ratios 2 to 4).
DISCUSSION
In this paper we describe a microanalytical method for
the quantitation of cardiolipin. The procedure involves
formation of a fluorescence derivative and its isolation
by a series of extractions. Quantitation was based on
fluorescence–HPLC in reference to an internal standard.
The standard compound, tetrastearoyl-cardiolipin, is a
synthetic analog that is uncommon in biological materials
(2–4, 21). This assay is specific for cardiolipin and is,
therefore, not suitable for the measurement of other
phospholipids. As expected for fluorescence analysis, the
assay was very sensitive. With the present array of HPLC
instruments, one nanomole of cardiolipin produced a sig-
nal of approximately 100 mV, putting the lower limit of
detection in the range of 10 to 20 pmol (corresponding to
1–2 mV on a total scale of 1.0 V). In practice, however, it
was difficult to measure such low quantities for two main
reasons: i ) the recovery of cardiolipin was only 18–20%;
Schlame et al. Muscle cardiolipin 1589

and ii) only 10–20% of a total sample could be injected
into the HPLC to avoid overloading with byproducts of
the chemical derivatization. Thus, no reliable signal was
obtained in biological samples with less than 0.5 nmol of
cardiolipin. The cardiolipin concentration, measured by
this method, correlated well with conventional phospho-
lipid measurements made by colorimetric determination
of phosphate after ashing (27).
The present method also allowed quantitation of indi-
vidual molecular species of cardiolipin, which were re-
solved by HPLC. A similar separation of molecular species
was achieved by a related technique, in which cardiolipin
was converted to the 2-benzoyl-cardiolipin dimethyl ester
(28). As in the previous method, there was a logarithmic
dependence of the HPLC retention time on the number
of double bonds in cardiolipins in which all acyl chains had
18 carbon atoms (Fig. 3). Presumably, a similar relationship
can be found for cardiolipins with other chain lengths but
appropriate molecular species were not available to test
this. The relationship between retention time and number
1590 Journal of Lipid Research Volume 40, 1999
Downloaded from www.jlr.org by guest, on June 7, 2013
Fig. 7. Electron micrographs of mitochondria from rat muscle
tissues. A: heart ventricle; B: heart auricular tissue; C: skeletal mus-
cle; D: esophagus smooth muscle; E: H9c2 myoblasts. The amplifi-
cation was 13,000.
of double bonds helped to identify species of muscle cardio-
lipins in routine analyses because they consisted almost en-
tirely of C18 chains. Compared to the previous method
(28), the new technique has a higher sensitivity and a
higher specificity, whereas the resolution of molecular spe-
cies is similar in the two methods. The increase of sensitivity
and specificity is due to the use of fluorescence detection as
well as the much improved sample work-up relying on spe-
cific extraction of derivatized cardiolipin.
In different muscle tissues, the concentration of cardio-
lipin (Fig. 4) corresponded roughly to the activity of cyto-
chrome oxidase (data not shown), confirming that cardio-
lipin is an indicator of mitochondrial content in the
tissue. The predominant molecular species in all muscles
examined were L 4-cardiolipin and L3O-cardiolipin. This
applied to heart muscle ( from rat and dog), skeletal mus-
cle (from humans and rat), and smooth muscle (from rat)
alike. However, there were significant differences in the
L 4/L3 O ratio in different types of muscle (Fig. 6). While
in rat heart ventricle and in rat skeletal muscle the L 4/L3O

Fig. 8. Cardiolipin content in skeletal muscle from patients with
mitochondrial disease and controls. Cardiolipin content was mea-
sured in skeletal muscle biopsies from adults (open symbols) and
children (closed symbols). Patients with mitochondrial disease had
cytochrome oxidase deficiency (squares), A3243G mutation in the
lipin. This, in turn, is consistent with the idea that the
function of L 4-cardiolipin is related to the density of pack-
ing in the cristae membrane. The cristae membrane has
to accommodate the multimeric complexes of oxidative
phosphorylation, plus, perhaps, an unknown number of
regulatory and assembly proteins, resulting in such high
protein concentration that it can be regarded as a lipopro-
tein (3).
Mitochondrial diseases are very hetereogeneous in clin-
ical manifestation, but are often associated with typical
morphological features in skeletal muscle (30, 31). In
adults, the most prominent of these features is the pres-
ence of “ragged-red fibers”, which are due to mitochon-
drial proliferation under the sarcolemma. Mitochondrial
proliferation seems to be a futile attempt to compensate
for single or multiple respiratory chain enzyme defects.
We found that cardiolipin, the only specific mitochondrial
phospholipid, is enriched in muscle biopsies from some
patients with a mitochondrial disease. This finding is con-
sistent with reactive proliferation of mitochondrial mem-
branes. However, increased tissue concentration of cardio-
lipin was not observed in all patients. The variability
within the disease group was much higher than in the

Downloaded from www.jlr.org by guest, on June 7, 2013

tRNALeu(UUR) gene resulting in MELAS (upward triangles), T8993G control group, probably reflecting variations between spe-
mutation of mitochondrial DNA resulting in Leigh syndrome
(downward triangle), and multiple deletions of mitochondrial DNA
cific defects or disease stages.
(diamond). Interestingly, we found a low proportion of L4-cardio-
lipin in three patients with mitochondrial DNA mutations.
As the L 4/L 3O ratio is related to cristae structure in rat
ratio was about 5, in smooth muscle it was only 2, and in muscle tissues, a low L 4 proportion in these patients may
the auricular appendage of rat heart it was about 1. Most be due to alterations of the inner mitochondrial mem-
striking was the difference between ventricle and auricu- brane. One possible mechanism is that dysfunctional
lar appendage as both were derived from heart tissue. mitochondria were unable to synthesize enough L 4-cardi-
This difference was also found in dog, although to a lesser olipin to match the growth of cristae membranes. L 4-car-
degree. In the H9c2 rat myoblast cell line, which is de- diolipin is probably the end-product of a fatty acid remod-
rived from embryonic heart tissue (29), the L 4/L 3O ratio eling cycle (32). Alternatively, L 4-cardiolipin could have
was even lower than in the auricular appendage. been degraded after oxidative damage, to which it is more
Mitochondria with a high L 4/L 3O ratio, such as those vulnerable than other cardiolipin species. This latter hy-
from heart ventricle and skeletal muscle, showed a con- pothesis is consistent with accumulating evidence that de-
densed conformation with stacked cristae (“muscle” type fects in the respiratory chain are accompanied by increased
mitochondria). In contrast, mitochondria from the auricular levels of free radicals.
appendage of rat heart, H9c2 myoblasts, or smooth mus- Manuscript received 29 March 1999 and in revised form 1 June 1999.
cle, which had a low L 4/L 3O ratio, were characterized by
an orthodox conformation with more relaxed cristae
REFERENCES
(“liver” type mitochondria). In rat liver the L 4/L 3O ratio
was 2.2 6 0.3 (n 5 3). These data suggest that tissues with 1. LeCocq, J., and C. E. Ballou. 1964. On the structure of cardiolipin.
more condensed mitochondria, reflecting a higher oxida- Biochemistry. 3: 976–980.
tive metabolism, need a higher proportion of L 4-cardio- 2. Hostetler, K. Y. 1982. Polyglycerophospholipids: phosphatidylglyc-
erol, diphosphatidylglycerol and bis(monoacylglycero)phosphate.
In Phospholipids, J. N. Hawthorne and G. B. Ansell, editors.
Elsevier Biomedical Press, Amsterdam. 215–261.
TABLE 1. Composition of cardiolipin in skeletal muscle 3. Daum, G. 1985. Lipids of mitochondria. Biochim. Biophys. Acta.
from patients with mitochondrial disease 822: 1–42.
4. Hoch, F. L. 1992. Cardiolipins and biomembrane function. Bio-
Composition chim. Biophys. Acta. 1113: 71–133.
Molecular COX Multiple 5. Schlame, M., and M. L. Greenberg. 1997. Cardiolipin synthase
Species Control Deficiency MELAS 3243 Deletions from yeast. Biochim. Biophys. Acta. 1348: 201–206.
6. Jiang, F., H. S. Rizavi, and M. L. Greenberg. 1997. Cardiolipin is not
L4 74.8 6 3.9 77.5 6 2.9 61.5 6 3.4 60.2 essential for the growth of Saccharomyces cerevisiae on fermentable
L 3O 17.4 6 2.1 13.3 6 1.7 20.9 6 3.5 22.5 and non-fermantable carbon sources. Mol. Microbiol. 26: 481–491.
L 2O2 5.8 6 1.1 4.7 6 0.6 9.3 6 2.4 10.4 7. Chang, S. C., P. N. Heacock, E. Mileykovskaya, D. R. Voelker, and
Other 1.9 6 1.2 4.5 6 1.7 8.3 6 2.6 6.9 W. Dowhan. 1998. Isolation and characterization of the gene
(cls1) encoding cardiolipin synthase in Saccharomyces cerevisiae. J.
COX, cytochrome oxidase. Biol. Chem. 273: 14933–14941.

Schlame et al. Muscle cardiolipin 1591

 8. Ohtsuka, T., M. Nishijima, K. Suzuki, and Y. Akamatsu. 1993. Mito-
chondrial dysfunction of a cultured Chinese hamster ovary cell
mutant deficient in cardiolipin. J. Biol. Chem. 268: 22914–22919.
9. Pangborn, M. C. 1942. Isolation and purification of serologi-
cally active phospholipid from beef heart. J. Biol. Chem. 143:
247 – 256.
10. Hostetler, K. Y. 1991. Effect of thyroxine on the activity of mito-
chondrial cardiolipin synthase in rat liver. Biochim. Biophys. Acta.
1086: 139–140.
11. Paradies, G., F. M. Ruggiero, P. Dinoi, G. Petrosillo, and E.
Quagliariello. 1993. Decreased cytochrome oxidase activity and
changes in phospholipids in heart mitochondria from hypothy-
roid rats. Arch. Biochem. Biophys. 307: 91–95.
12. Schlame, M., and K. Y. Hostetler. 1997. Cardiolipin synthase
from mammalian mitochondria. Biochim. Biophys. Acta. 1348:
207 – 213.
13. Floridi, A., R. Pulselli, F. P. Gentile, R. Barbieri, and M. Benassi.
1994. Rhein enhances the effect of adriamycin on mitochondrial
respiration by increasing antibiotic-membrane interaction. Bio-
chem. Pharmacol. 47: 1781–1788.
14. Iwase, H., T. Takatori, M. Nagao, K. Iwadate, and M. Nakajima.
1996. Monoepoxide production from linoleic acid by cytochrome
c in the presence of cardiolipin. Biochem. Biophys. Res. Commun.
222: 83–89.
15. Rossen, R. D., L. H. Michael, H. K. Hawkins, K. Youker, W. J.
Dreyer, R. E. Baughn, and M. L. Entman. 1994. Cardiolipin-
protein complexes and initiation of complement activation after
coronary artery occlusion. Circ. Res. 75: 546–555.
16. Paradies, G., F. M. Ruggiero, G. Petrosillo, and E. Quagliariello.
1997. Age-dependent decline in the cytochrome c oxidase activity
in rat heart mitochondria: role of cardiolipin. FEBS Lett. 406: 136–
138.
17. Petit, J. M., A. Maftah, M. H. Ratinaud, and R. Julien. 1992. 10-N-
nonyl acridine orange interacts with cardiolipin and allows the
quantification of this phospholipid in isolated mitochondria. Eur.
J. Biochem. 209: 267–273.
18. Paradies, G., and F. M. Ruggiero. 1989. Decreased activity of the
pyruvate translocator and changes in the lipid composition in
heart mitochondria from hypothyroid rats. Arch. Biochem. Biophys.
269: 595–602.
1592 Journal of Lipid Research Volume 40, 1999
View publication stats
19. Paradies G., and F. M. Ruggiero. 1990. Stimulation of phosphate
transport in rat liver mitochondria by thyroid hormones. Biochim.
Biophys. Acta. 1019: 133–136.
20. Cao, S. G., P. Cheng, A. Angel, and G. M. Hatch. 1995. Thyroxine
stimulates phosphatidylglycerophosphate synthase activity in rat
heart mitochondria. Biochim. Biophys. Acta. 1256: 241–244.
21. Schlame, M., S. Brody, and K. Y. Hostetler. 1993. Mitochondrial car-
diolipin in diverse eukaryotes. Comparison of biosynthetic reactions
and molecular acyl species. Eur. J. Biochem. 212: 727–735.
22. Blumenthal, D. T., S. Shanske, S. S. Schochet, F. M. Santorelli,
S. DiMauro, M. Jaynesm, and J. Bodensteiner. 1998. Myoclonus
epilepsy with ragged red fibers and multiple mtDNA deletions.
Neurology. 50: 524–525.
23. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid ex-
traction and purification. Can. J. Biochem. Physiol. 37: 911–917.
24. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951.
Protein measurement with the Folin phenol reagent. J. Biol. Chem.
193: 265–275.
25. Patton, G. M., S. Cann, H. Brunengraber, and J. M. Lowenstein.
1981. Separation of methyl esters of fatty acids by gas chromatog-
raphy on capillary columns, including the separation of deuter-
ated from non-deuterated fatty acids. Methods Enzymol. 72: 8–20.
26. Srere, P. A. 1969. Citrate synthase. Methods Enzymol. 13: 3–11.
27. Zhou, X., and G. Arthur. 1992. Improved procedures for the deter-
mination of lipid phosphorus by malachite green. J. Lipid Res. 33:
1233–1236.
28. Schlame, M., and D. Otten. 1991. Analysis of cardiolipin molecular
species by high-performance liquid chromatography of its deriva-
tive 1,3-bisphosphatidyl-2-benzoyl-sn-glycerol dimethyl ester. Anal.
Biochem. 195: 290–295.
Downloaded from www.jlr.org by guest, on June 7, 2013
29. Kimes, B. W., and B. L. Brandt. 1976. Properties of a clonal muscle
cell line from rat heart. Exp. Cell Res. 98: 367–381.
30. DiMauro, S., M. Hirano, E. Bonilla, and D. C. De Vivo. 1996. The
mitochondrial disorders. In Principles of Child Neurology. O.
Berg, editor. McGraw-Hill, New York. 1201–1232.
31. Harding, A. E. 1991. Neurological disease and mitochondrial
genes. Trends Neurol. Sci. 14: 132–138.
32. Schlame, M., and B. Rüstow. 1990. Lysocardiolipin formation and
reacylation in isolated rat liver mitochondria. Biochem. J. 272: 589–
595.